arXiv cs.AI
·
18 hours ago
A new reinforcement learning algorithm called KGRL combines symbolic knowledge from Datalog rule bases with gradient-guided parameter optimization to improve training efficiency in tasks requiring both discrete action selection and continuous parameter tuning. The method prunes infeasible actions and constrains parameter spaces based on domain knowledge, achieving superior sample efficiency and return compared to existing baselines. This approach enables agents to learn constraint-aware decisions while providing interpretable explanations of action selection and parameter constraints.
arXiv cs.AI
·
18 hours ago
Researchers developed a neuro-symbolic verification system that wraps language models to generate valid twelve-tone musical compositions, checking outputs against formal constraints before delivery. The system improved constraint-checked delivery from 13.3% to 48.1% across 40 controlled tasks with four paired models, with the harness abstaining on 51.9% of runs that failed verification. Expert evaluation showed preference for the harness-generated compositions over raw model output in adherence, perceived legality, and overall musical quality.
arXiv cs.AI
·
18 hours ago
Researchers introduced NeSy-Route, a neuro-symbolic benchmark for evaluating multimodal large language models on constrained route planning tasks in remote sensing applications. The benchmark contains 10,821 route-planning samples generated through an automated framework combining semantic masks with heuristic search to produce provably optimal solutions. Existing MLLMs demonstrated significant deficiencies in perception and planning capabilities when evaluated using the new three-level hierarchical assessment protocol.
arXiv cs.AI
·
18 hours ago
Researchers extended neuro-symbolic AI based on Belnap's intensional first-order logic by incorporating probabilistic reasoning for unknown sentences using Nilsson's probability structure. The approach introduces global and local symmetry transformations to preserve knowledge while computing probability density functions through neural networks using maximum entropy principles. This enables AGI systems to combine neural learning with symbolic reasoning while handling uncertainty in logical reasoning tasks.
